Clinical Science (2006) 111, (Printed in Great Britain) doi: /cs

Transcription

1 Clinical Science (2006) 111, (Printed in Great Britain) doi: /cs Association study of ACE2 (angiotensin I-converting enzyme 2) gene polymorphisms with coronary heart disease and myocardial infarction in a Chinese Han population Wei YANG, Wentao HUANG, ShaoyongSU,BiaoLI, Weiyan ZHAO, Shufeng CHEN and Dongfeng GU Division of Population Genetics and Prevention, Cardiovascular Institute, Fu Wai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 167 Beilishi Road, Beijing , People s Republic of China, National Human Genome Center at Beijing, North Yongchang Road 3-707, Beijing , People s Republic of China, and Institute of Biophysics, Chinese Academy of Sciences, Datun Road 15, Chaoyang District, Beijing , People s Republic of China A B S T R A C T Results are accumulating that ACE2 (angiotensin I-converting enzyme 2) might act as a protective protein for cardiovascular diseases; however, only a few studies in human populations have been carried out. This prompted us to perform a case-control study to investigate the relationship of ACE2 polymorphisms with CHD (coronary heart disease) and MI (myocardial infarction). Three single nucleotide polymorphisms in the ACE2 gene (1075A/G, 8790A/G and 16854G/C) were genotyped by PCR-RFLP (restriction-fragment-length polymorphism) in 811 patients with CHD (of which 508 were patients with MI) and 905 normal controls in a Chinese population. The polymorphisms were in linkage disequilibrium (r 2 = ). Analyses were conducted by gender, because the ACE2 gene is on the X chromosome. In females, an association was detected with MI for 1075A/G (P = 0.026; odds ratio = 1.98) and 16854G/C (P = 0.028; odds ratio = 1.97) in recessive models after adjusting for covariates. In male subjects, two haplotypes (AAG and GGC) were common in frequency. In male subjects not consuming alcohol, the haplotype GGC was associated with a 1.76-fold risk of CHD [95 % CI (confidence interval), ; P = 0.007] and a 1.77-fold risk of MI (95 % CI, ; P = 0.015) with environmental factors adjusted, when compared with the most common haplotype AAG. In conclusion, the results of the present study indicate that common genetic variants in the ACE2 gene might impact on MI in females, and may possibly interact with alcohol consumption to affect the risk of CHD and MI in Chinese males. Key words: association study, angiotensin I-converting enzyme 2 (ACE2), coronary heart disease, gender, myocardial infarction, single nucleotide polymorphism. Abbreviations: Ang, angiotensin; ACE, Ang I-converting enzyme; BMI, body mass index; BP, blood pressure; CHD, coronary heart disease; CI, confidence interval; DBP, diastolic BP; HDL-C, high-density lipoprotein-cholesterol; HWE, Hardy Weinberg equilibrium; LD, linkage disequilibrium; LDL-C, low-density lipoprotein-cholesterol; MAF, minor allele frequency; MI, myocardial infarction; OR, odds ratio; RAS, renin angiotensin system; SBP, systolic BP; SNP, single nucleotide polymorphism; TC, total cholesterol. Correspondence: Professor Dongfeng Gu, at the Division of Population Genetics and Prevention, Cardiovascular Institute, Fu Wai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 167 Beilishi Road, Beijing , People s Republic of China ( gudf@yahoo.com).

2 334 W. Yang and others INTRODUCTION CHD (coronary heart disease), including its most severe complication MI (myocardial infarction), is a leading cause of death worldwide. It is believed to be caused by many genetic and environmental factors as well as their interactions. The prevailing view is that CHD/MI is a disease of inflammation and immune responses [1], and atherosclerosis underlies the major pathophysiological mechanisms [2]. Genes in the RAS (renin angiotensin system) have been widely used for association studies of cardiovascular diseases, in particular the ACE [Ang (angiotensin) I-converting enzyme 1] gene [3]. ACE cleaves the inactive decameric peptide Ang I into the active octomer Ang II, which is the biological effector of the RAS system. Ang II is widely related to pro-atherosclerotic and inflammatory actions [4]. ACE2 (Ang I-converting enzyme 2), a homologue of ACE, has recently been identified and considered to act as an anti-ras element [5]. It is predominantly expressed in vascular endothelial cells of the kidney and heart. It cleaves a single residue from Ang I to generate Ang-(1 9), and metabolizes Ang II to generate Ang-(1 7) [6,7], which is a potent vasodilator and antitrophic peptide that has generated considerable interest during the past decade [8,9]. Thus ACE2 may counteract the function of ACE either by converting Ang II into Ang- (1 7) or by competing with ACE for the same substrate Ang I [10]. Besides the Angs, it also hydrolyses other peptide substrates. For instance, it can degrade bioactive opioid peptides and generate apelin peptides and kinin metabolites that activate APJ (Ang-receptor-like 1) and B 1 receptors respectively [11]. TheACE2geneislocatedontheXchromosome (Xp22) with 19 exons plus a newly identified 5 -untranslated exon [6,7,12]. It has been mapped to a QTL (quantitative trait locus) associated with hypertension in three rat models of high BP (blood pressure) [10]. In ACE2-deficient mice, there is an increase in Ang II levels, even though no apparent effect on BP was observed. A reduction in cardiac contractility and a significant decrease in aortic and ventricular pressure was also observed in these mice [10]. A potential role for ACE2 in contributing to the cardioprotective effect of Ang-(1 7) was demonstrated when long-term infusions of Ang-(1 7) reversed cardiac dysfunction and restored vascular endothelial responses in animals after MI [13]. Burrell et al. [14] found that MI increased ACE2 expression in rats and humans. These data suggested a relationship between the ACE2 gene and CHD. In contrast with the rapid accumulation of findings about the important role of the ACE2 gene in cardiovascular diseases from animal-model-based experiments, only a few studies have been carried out to investigate the gene in human populations. A study in Australians of white Anglo Celtic origin reported no association of ACE2 gene polymorphisms with essential hypertension [15], and Chen et al. [16], in a meeting abstract, suggested associations between two polymorphisms of this gene and a number of cardiovascular end points in a Caucasian population. A recent study in a German population [17] found a haplotype consisting of four SNPs (single nucleotide polymorphisms) spanning the gene associated with higher risk of left ventricular mass index, septal wall thickness and left ventricular hypertrophy. As far as we know, whether the ACE2 gene is associated with human CHD/MI has not been directly addressed yet. This led us to conduct an association study between ACE2 gene polymorphisms and human CHD/MI. MATERIALS AND METHODS Subjects The enrolment criteria of the CHD cases and controls for the Beijing Atherosclerosis Study have been reported in detail previously [18]. We recruited 811 patients with CHD from patients hospitalized at the Fu Wai Hospital and Cardiovascular Institute (Beijing, China) between October 1997 and December Eligible patients were those who survived an acute MI or were documented by coronary angiography to have evidence of at least a 70 % stenosis in a major epicardial artery. Patients with congenital heart disease, cardiomyopathy, valvular disease and renal or hepatic disease were excluded. As a subset of the patients with CHD, the MI group consisted of 508 patients. A total of 905 control subjects were randomly selected from individuals participating in a community-based survey of cardiovascular risk factors. The controls were judged to be free of ischaemic changes by ECG, to have no symptoms of chest pain and to be free of CHD by medical history, the Rose questionnaire [19] and clinical examination. This study was approved by the local Research Ethnics Committee, and all subjects gave written informed consent. A set of questionnaires was completed that included details of medical history, family history of CHD, drug intake, cigarette smoking and alcohol consumption. BP, weight, height, waistline and hip circumference were recorded, and BMI (body mass index) was calculated. Hypertension was defined as SBP (systolic BP) 140 mmhg and/or DBP (diastolic BP) 90 mmhg, a history of hypertension, or the patient taking antihypertension medication. Venous blood was drawn from all subjects after an overnight fast. Blood, serum and plasma were separated immediately and stored at 70 C. SNP selection and genotyping Three SNPs, 1075A/G (rs in intron 1), 8790A/G (rs in intron 3) and 16854G/C (rs in

3 ACE2 gene variants and CHD 335 intron 7), were selected as proxies to study ACE2 polymorphisms. This was justified by the strong LD (linkage disequilibrium) throughout the gene which has been suggested in various studies [15 17,20] (details in the Discussion section). All three SNPs had been involved in other studies [15,16,20]. Genomic DNA was isolated from white blood cells using the phenol/chloroform method and was stored in 400 µl of TE [10mM Tris/HCl and 1mM EDTA (ph 8.0)]. The three SNPs were genotyped using standard PCR- RFLP (restriction-fragment-length polymorphism) protocols. Primers and conditions for amplification are available from the authors upon request. Quality control was performed by re-genotyping all three SNP in 98 subjects randomly selected from the control group. The replication of genotyping was performed by another team member. A comparison between the two showed that the discrepancy was 1.36 %, with 95 % CI (confidence interval) ranging from %. Statistical analyses Analyses were taken separately for the three SNPs and were followed by haplotype analyses. In consideration of the different number of copies of the X chromosome in males and females, separate analyses were taken for each subgroup. Departure from the HWE (Hardy Weinberg equilibrium) was tested among females by the χ 2 test. LD was estimated using GOLD software [21]. For single locus analyses, genotype and allele frequency distributions across healthy individuals and patients were tested by χ 2 test, and for female genotypes, association was analysed under additive, dominant and recessive models respectively. Furthermore, covariates, including age, BMI, smoking, alcohol consumption, history of hypertension, triacylglycerols (triglycerides), TC (total cholesterol), HDL-C (high-density lipoprotein-cholesterol) and glucose were adjusted by logistic regression to investigate the independent role of each polymorphism. Male haplotypes were readily available, and we used the EM algorithm to account for missing genotypes (13 missed for 16854G/C). Both χ 2 test and backward stepwise logistic regression were used to assess the association between the male haplotypes and CHD/MI. In females, the association of haplotypes with diseases was analysed using the Haplo.score approach, as outlined by Schaid et al. [22]. This approach was implemented in the haplo.stats software developed using the R language ( The method helps with the inference of haplotypes and accounts for haplotype ambiguity. It can be used to adjust for covariates. Estimated haplotype pairs were also taken for diplotype analysis. We analysed further the interaction between the genetic effects (genotypes and haplotypes) and effects of alcohol consumption, smoking and BMI by introducing the corresponding product terms into the logistic regression models. General calculations, including χ 2 tests and logistic regression, were performed using the Stata software (Stata/SE 8.0 for windows). We used P < 0.05 to define statistical significance. In backward stepwise logistic regression, we used significant levels of 0.2 for removal from the model and of 0.05 for addition into the model. RESULTS The demographic details of the patients and controls are given in Table 1. Compared with the control group, the CHD group had more patients with hypertension, higher mean BMI and SBP, higher levels of serum triacylglycerols, LDL-C (low-density lipoprotein-cholesterol) and fasting glucose, and lower HDL-C levels. However, DBP was significantly lower in patients with CHD, which could be a result of medication taken after they were diagnosed, as 11.7 % of the controls and 44.5 % of the patients were taking antihypertension medication. Characteristics of the MI group of patients were similar with those of the CHD group, except that TC and LDL-C levels were not significantly different compared with the controls, and that the MI group had larger proportions of smokers and alcohol consumers. Genotype and allele frequencies for each polymorphism in male and female subjects are shown separately in Tables 2 and 3 respectively. MAF (minor allele frequency) was higher in the female controls compared with the male controls for each of the SNPs, but this difference was not statistically significant. HWE could be tested in only the female population and showed no significant deviation for all of the three sites in the control group and in patients with CHD. However, in female patients in the MI group, HWE was not in accordance with expectation for any of the SNPs. Pairwise LD tests indicated that all of the three SNPs investigated were in strong LD with each other (D = , P < ; r 2 = ). None of the polymorphisms had any association with CHD or MI in male patients (Table 2), and neither did they have any association with female CHD in additive, dominant or recessive models (Table 3). Calculating the ORs (odds ratios) for female patients with CHD and MI suggested a recessive model effect, and association tests showed that carriers of the 1075AA and 16854GG genotypes were at higher risks of MI (P = and respectively; OR = 1.74 and 1.67 respectively; Table 3). Subsequent stepwise logistic regression to adjust for covariates revealed that the 1075A/G and 16854G/C polymorphisms were associated with MI in female patients (P = and 0.028; OR, 1.98 and 1.97 respectively) in a recessive mode of inheritance. The same trend was found with 8790A/G, but this was only marginally significant.

4 336 W. Yang and others Table 1 Characteristics of the patients with CHD and MI compared with controls For continuous characteristics, value are means + S.D.; for discrete characteristics, values are the number of observations. The MI group are a subset of the CHD group. Not all measurements were available for all participants, but a maximum of 11 values were absent at the most for each characteristic. P < 0.05 compared with controls. Patients with Characteristic Controls CHD MI Sex (male/female) 665/ / /92 Age (years) BMI (kg/m 2 ) TC (mg/dl) HDL-C (mg/dl) Triacylglycerols (mg/dl) LDL-C (mg/dl) Glucose (mg/dl) SBP (mmhg) DBP (mmhg) Hypertension (yes/no) 299/ / /211 Smokers (yes/no) 522/ / /167 Alcohol consumers (yes/no) 385/ / /260 Table 2 Genotype and allele frequencies of ACE2 polymorphism in male subjects P values are compared with controls G/C genotype was missing for 13 male subjects. Patients with SNPs Alleles Controls CHD P value MI P value 1075A/G 8790A/G 16854G/C A/G Count (n) 381/ / /195 MAF (%) A/G Count (n) 381/ / /196 MAF (%) C/G Count (n) 378/ / /194 MAF (%) Total (n) In male subjects, two haplotypes, HAP1 (AAG) and HAP2 (GGC), were common in frequency. The other five haplotypes in all constituted < 5 % of the total number. Frequencies of haplotypes in males were not significantly different between controls and patients with CHD or MI whether environmental covariates were adjusted for or not. In females, haplotype was associated with neither CHD nor MI, but diplotype analysis revealed the carriers of two copies of HAP1 were at higher risk of MI (OR = 1.81, P = 0.022) compared with other diplotypes, as expected. When environmental factors were adjusted for, this association was more significant (P = and OR = 2.1 for diplotype HAP1/HAP1 compared with the others). Alcohol consumption might interact with the male SNPs and haplotypes, as the corresponding product terms in logistic regression analysis had marginal significant contributions (P = ) for CHD. This association was investigated further in subgroups of male subjects. In males not consuming alcohol, HAP2 carriers had a significantly increased risk of CHD [OR = 1.76 (95 % CI, ); P = 0.007] and MI [OR = 1.77 (95 % CI, ); P = 0.015] after adjusting for environmental factors when compared

5 ACE2 gene variants and CHD 337 Table 3 Genotype and allele frequencies of the ACE2 polymorphisms in female subjects P values are compared with controls G/C genotype was missing for three female subjects. Patients with SNPs Alleles Controls CHD P value (r/d ) MI P value (r/d ) 1075A/G 8790A/G 16854G/C AA/AG/GG Count (n) 69/112/59 59/73/43 38/30/24 Frequency (%) 28.8/46.7/ /41.7/ (0.31/0.98) 41.3/32.6/ (0.028/0.78) MAF (%) AA/AG/GG n 68/116/56 57/75/43 36/33/23 Frequency (%) 28.3/48.3/ /42.9/ (0.39/0.75) 39.1/35.9/ (0.058/0.79) MAF (%) GG/GC/CC n 68/116/53 58/76/41 37/33/22 Frequency (%) 28.7/48.9/ /43.4/ (0.37/0.78) 40.2/35.9/ (0.044/0.76) MAF (%) Total (n) r/d, P values under recessive/dominant models. with HAP1. However, in males consuming alcohol, the distribution of haplotypes and genotypes between cases and controls was not significantly different (P = ). Table 4 Comparison of MAFs in different studies Mainland Chinese Hong Kong Chinese Australian Caucasians Male Female Male Female Male Female DISCUSSION In contrast with the rapid accumulation of findings on the important role of the ACE2 gene in cardiovascular diseases from animal-model-based experiments, only a few studies have been carried out to investigate the gene in human populations. Benjafield et al. [15] found no association between ACE2 polymorphisms and hypertension, and stated that the role of ACE2 in CHD remained to be examined. Chen et al. [16] reported in a meeting abstract that two SNPs in the ACE2 locus were associated with a number of cardiovascular end points, although they did not provide any specific detailed information on the SNPs or the disease of interest. A recent study in a German population [17] found a haplotype consisting of four SNPs spanning the ACE2 gene that was associated with a higher risk of left ventricular mass index, septal wall thickness and left ventricular hypertrophy in males and a similar, but less, pronounced trend in females. In the present study, two sites of all three of the SNPs studied and the HAP1/HAP1 diplotype were associated with MI in females in recessive models, and the haplotypes were associated with both CHD and MI in males not consuming alcohol. Thus the findings of the present study do not contradict these previous observations in humans. n MAF (%) 1075A/G A/G G/C Data from the present study Data from a study of severe acute respiratory syndrome in Hong Kong Chinese subjects [20]. Data from Benjafield et al. [15] The three SNPs have also been genotyped in previous studies [15,20] and thus the MAFs were able to be compared with those from the present study. As shown in Table 4, MAFs for 8790A/G and 16854G/C in our present study were approximately equal to that in a Hong Kong Chinese population [20], but these were much lower for 1075A/G and 8790A/G in an Australian Caucasian population of Anglo Celtic origin in both males and females [15]. This suggests a consistency within Chinese populations, but a large race-wise difference between Chinese and Caucasian populations. The HapMap website ( provides information on LD across the whole genome in four different populations, including a Chinese Han population. However, the data quality for the ACE2 gene are questionable. Only 11 out of 71 genotyped SNPs in this gene were considered valid by Haploview

6 338 W. Yang and others software when analysed using default configurations. Two of the SNPs investigated in the present study, 1075A/ G and 16854G/C, were contained within the HapMap listing, but both had MAFs of 0 in a Chinese population. However, these SNPs are obviously common polymorphisms in both Chinese and Caucasian populations, as they have been reported in the previous studies mentioned above. In the HapMap project, more than 1 million SNPs were genotyped from only 45 Chinese, 45 Japanese, 90 African and 90 European subjects, and this might partly explain the situation. The ACE2 gene appears to be in a single haplotype block, even though no information about this is available from the HapMap database. In our present study, the three SNPs, 1075A/G (intron 1), 8790A/G (intron 3) and 16854G/C (intron 7), were in strong LD (D = ; r 2 = ). Benjafield et al. [15] showed that 1075A/G (intron 1), 8790A/G (intron 3), rs (intron 11) and rs (intron 16) were also in LD (D = ; P = ). Chen et al. [16] reported that an SNP in intron 3 and an inserting/deletion SNP in intron 9 were in strong LD (D = 0.93; P < 0.001). In a recent study in a German population [17], four SNPs in introns 3, 11, 14 and 16 respectively, also had high pairwise LD. Even though these studies were not in the same populations, strong LD across the gene was supported. This implies that studying a few SNPs in this LD block will provide enough information for all other SNPs. As an example of this, in the present study, any of the three SNPs provides almost the same information. In females, carriers of the 1075AA and 16854GG genotypes were at higher risk of MI in recessive models, which was confirmed by subsequent logistic regression analyses. Co-incidentally, only the genotypes among female patients with MI were in discordance with HWE for each of the SNPs. Lee [23] has proposed a method for searching disease-susceptibility loci by testing for Hardy Weinberg disequilibrium in a gene bank of affected individuals. Deviation from HWE among cases, while HWE does not deviate in controls, could reflect an underlying association. Thus the co-incidence might be explained by a real association. Diplotype analysis provided further support for this suggestion, but we cannot exclude the possibility of false positives, due to the low sample size in our female population (240 controls and 92 patients with MI). Existence of gene environment interactions has been implicated as a complicating factor in discovering genetic models of complex human diseases. In the present study, we found a possible interaction between alcohol consumption and ACE2 gene polymorphisms in male subjects. The effect of alcohol consumption on CHD has been studied and, in a recent review, Hill [24] argued that mild-to-moderate consumption of alcohol tends to play a protective role in the pathogenesis of CHD, but risk is promoted for heavy consumers. The effect is also gender-specific. In the present study, alcohol consumption was defined as using alcohol at least once a week for at least 6 months. Thus quantitative analysis of its effect is not feasible. It should also be noted that the data are self-reported and have to be interpreted with some caution. Additional data will be required to confirm this potential interaction. One possible approach is to include both ADH1C (formerly called ADH3) and ACE2 genes in a similar study. ADH1C encodes an alcohol dehydrogenase enzyme that mediates the initial breakdown of alcohol, and there are two common genetic variants of the gene that breakdown ethanol at different speeds. Thus existence of an interaction between ADH1C and ACE2 polymorphisms may confirm the interaction reported in the present study. In our present study, one of these SNPs would have been sufficient to study when the strong LD among each pair of the SNPs is considered. Even though predetermined at the stage of study design, we proceeded with the analysis of three SNPs and haplotype analysis in the CHD and MI groups, and also performed analysis of interactions. It is encouraging that every single SNP and haplotype had an almost identical result, which reduces the possibility of making false judgments; however, it should be noted that the P values were not corrected for multiple testing and further work is therefore necessary. HAP1 was associated with an increased risk of MI in females, whereas HAP2 was associated with increased risk of CHD and MI in male subjects consuming alcohol. Gender-specific effects of the RAS on cardiovascular disease like this have been observed and studied widely. Both androgens and oestrogens play a role in modulating the effects of members of the RAS, and numerous studies have demonstrated that genetic polymorphisms in ACE, AGT (angiotensinogen) or AGTR1 (Ang II type 1 receptor) genes display more profound effects in men than in women [25,26]. As a newer member of the RAS, ACE2 might also be modulated by sex hormones to exert its effect. In the presence of interactions, the observed effect of ACE2 might be distorted when ignoring the context of other factors where it takes effect, even though it still suggests an association between disease and genotypes. This highlights the need for further gene environment interaction studies of the ACE2 gene in CHD using a systematic approach. How ACE2 polymorphisms affect the pathogenesis of CHD and MI remains largely unclear. ACE2 is obviously multifunctional given the observation that it has been widely used to study severe acute respiratory syndrome [20]. It might impact on CHD by altering the levels of Ang II and Ang-(1 7). Ang II mediates endothelial dysfunction via increased oxidative stress in cell culture, animal models and human subjects. Alternatively, Ang-(1 7), besides causing vasodilation, counteracts the growth-promoting and fibrotic effects of Ang II. However, novel roles for ACE2 in hydrolysing several

7 ACE2 gene variants and CHD 339 other peptides, such as the apelins, opioids and kinin metabolites raises the possibility that peptide systems other than Ang and its derivatives may also have an important role in regulating cardiovascular function [5,11]. Several limitations of our present study should be considered. First, we selected only three intronic SNPs in the kb ACE2 gene. This is partly because SNPs in exons are both scarce in terms of numbers and are also at low frequencies based on information from dbsnp (Single Nucleotide Polymorphism database; and other sources. Systematic selection of SNPs should be made considering LD throughout the gene, and variants in conservative regions should also be accounted for. However, since the whole gene is likely to be located in one LD block, the problem might not be critical. Secondly, the problem of a deficiency in sample size should be addressed when analysing gene environment interactions. In our present study, the sample size of the female subjects was too small to afford stratification, and the male sample was not sufficient to ensure adequate power after stratification. In fact, there are 282 controls, 274 patients with CHD (including 178 patients with MI) in males not consuming alcohol. When comparison of the frequencies of HAP2 were made between the 282 male controls and 178 male patients with MI without adjustment, power to detect a true association was as low as The problem of subgrouping for interaction analyses has been discussed elsewhere [27]. The demand for sample size grows dramatically when a confounding interactive factor is taken into account. Larger samples should be considered in subsequent studies. Finally, if alcohol consumption is quantitatively assessed, results could be more exact, as both excessive consumption and abstinence is deleterious in this scenario. In conclusion, the present study indicates (i) the association of common genetic variants in the ACE2 gene with MI in females, and (ii) a potential interaction between these variants and alcohol consumption to affect the risk of CHD and MI in Chinese males. Further studies with large sample sizes and based on different population backgrounds are needed to elucidate this relationship. ACKNOWLEDGMENTS This study was funded by grants 2002BA711A05, 2002BA711A08, and 2002BA711A10 from The National Tenth Five-Year Plan Key Programs from the Ministry of Science and Technology of The People s Republic of China, H Biomedical Project from the Council of Science and Technology, Beijing, and grant from the National Natural Science Foundation of China. REFERENCES 1 Wang, Q. (2005) Molecular genetics of coronary artery disease. Curr. Opin. Cardiol. 20, Hansson, G. K. (2005) Inflammation, atherosclerosis, and coronary artery disease. N. Engl. J. Med. 352, Unger, T. (2002) The role of the renin-angiotensin system in the development of cardiovascular disease. Am. J. Cardiol. 89, 3A 9A 4 O Keefe, J. H., Lurk, J. T., Kahatapitiya, R. C. and Haskin, J. A. (2003) The renin-angiotensin-aldosterone system as a target in coronary disease. Curr. Atheroscler. Rep. 5, Kurdi, M., De Mello, W. C. and Booz, G. W. (2005) Working outside the system: an update on the unconventional behavior of the renin-angiotensin system components. Int. J. Biochem. Cell. Biol. 37, Tipnis, S. R., Hooper, N. M., Hyde, R., Karran, E., Christie, G. and Turner, A. J. (2000) A human homolog of angiotensin-converting enzyme. Cloning and functional expression as a captopril-insensitive carboxypeptidase. J. Biol. Chem. 275, Donoghue, M., Hsieh, F., Baronas, E. et al. (2000) Acton S. A novel angiotensin-converting enzyme-related carboxypeptidase (ACE2) converts angiotensin I to angiotensin 1 9. Circ. Res. 87, E1 E9 8 Li, N., Zimpelmann, J., Cheng, K., Wilkins, J. A. and Burns, K. D. (2005) The role of angiotensin converting enzyme 2 in the generation of angiotensin 1 7 by rat proximal tubules. Am. J. Physiol. Renal Physiol. 288, F353 F362 9 Clark, M. A., Diz, D. I. and Tallant, E. A. (2001) Angiotensin-(1 7) downregulates the angiotensin II type 1 receptor in vascular smooth muscle cells. Hypertension 37, Crackower, M. A., Sarao, R., Oudit, G. Y. et al. (2002) Angiotensin-converting enzyme 2 is an essential regulator of heart function. Nature 417, Oudit, G. Y., Crackower, M. A., Backx, P. H. and Penninger, J. M. (2003) The role of ACE2 in cardiovascular physiology. Trends Cardiovasc. Med. 13, Itoyama, S., Keicho, N., Hijikata, M. et al. (2005) Identification of an alternative 5 -untranslated exon and new polymorphisms of angiotensin-converting enzyme 2 gene: lack of association with SARS in the Vietnamese population. Am. J. Med. Genet. A 136, Loot, A. E., Roks, A. J., Henning R.H., A. J. et al. (2002) Angiotensin-(1 7) attenuates the development of heart failure after myocardial infarction in rats. Circulation 105, Burrell, L. M., Risvanis, J., Kubota, E. et al. (2005) Myocardial infarction increases ACE2 expression in rat and humans. Eur. Heart J. 26, Benjafield, A. V., Wang, W. Y. and Morris, B. J. (2004) No association of angiotensin-converting enzyme 2 gene (ACE2) polymorphisms with essential hypertension. Am. J. Hypertens. 17, Chen, A. A., Barnes, G., Foti, A. et al. (2002) Two single nucleotide polymorphisms in the ACE2 locus are associated with cardiovascular disease. Gen. Epidem. 23, Lieb, W., Graf, J., Gotz, A. et al. (2006) Association of angiotensin-converting enzyme 2 (ACE2) gene polymorphisms with parameters of left ventricular hypertrophy in men. J. Mol. Med. 84, Zhou, X., Huang, J., Chen, J. et al. (2004) Thrombospondin-4 A387P polymorphism is not associated with coronary artery disease and myocardial infarction in the Chinese Han population. Clin. Sci. 106, Rose, G. A. and Blackburn, H. (1968) Cardiovascular survey methods. Monogr. Ser. World Health Organ. 56, Chiu, R. W., Tang, N. L., Hui, D. S. et al. (2004) ACE2 gene polymorphisms do not affect outcome of severe acute respiratory syndrome. Clin. Chem. 50, Abecasis, G. R. and Cookson, W. O. (2000) GOLD: graphical overview of linkage disequilibrium. Bioinformatics 16,

8 340 W. Yang and others 22 Schaid, D. J., Rowland, C. M., Tines, D. E., Jacobson, R. M. and Poland, G. A. (2002) Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am. J. Hum. Genet. 70, Lee, W. C. (2003) Searching for disease-susceptibility loci by testing for Hardy-Weinberg disequilibrium in a gene bank of affected individuals. Am. J. Epidemiol. 158, Hill, J. A. (2005) In vino veritas: alcohol and heart disease. Am. J. Med. Sci. 329, Marcus, F., Andrea, B. and Heribert, S. (2002) Renin angiotensin system and gender differences in the cardiovascular system. Cardiovasc. Res. 53, Jane, F. R. (2001) Gender differences in the regulation of blood pressure. Hypertension 37, Hunter, D. J. (2005) Gene environment interactions in human diseases. Nat. Rev. Genet. 6, Received 26 January 2006/19 June 2006; accepted 6 July 2006 Published as Immediate Publication 6 July 2006, doi: /cs